Author Topic: Is there something inherently noisy in this design?  (Read 2374 times)

Bill Mountain

Is there something inherently noisy in this design?
« on: May 21, 2012, 01:11:04 PM »
I know that inverting opamp stages are generally noisier than non-inverting opamp stages and cmos inverters are always noiser than their opamp cousins but when trying this out on the breadboard there was far more hiss than normal.  I was attempting to make a high input impedance booster that went from -6dB to +30dB and I thought I had it with this design but I was forced to go with a more traditional boost stages to keep the noise down.  I have always thought about trying this again and I figured I’d ask you all how this could be made a tad bit less noisy.  Plus, I noticed that the there was a low pass effect based on the relationship between the 470k and the 3.9k.  I thought I could exploit this to tame some high end but I’m not sure how to calculate it.  If I put a cap to ground (smallish - about 100pf) between the 1M and 250kA pot then the LPF would get really steep so that was worth trying to figure out as well.  The 10uf wasn’t necessary when I simulated the circuit but I needed it on the breadboard most likely due to running on single supply.  Lower values here would have some noticeable high pass effect so I thought of putting a cap selector here for more control.



Thoughts?
« Last Edit: May 21, 2012, 01:14:37 PM by Bill Mountain »

Gurner

Re: Is there something inherently noisy in this design?
« Reply #1 on: May 21, 2012, 01:17:45 PM »
The 10uf wasn’t necessary when I simulated the circuit but I needed it on the breadboard most likely due to running on single supply.

If you are running on a single supply, then that 1st opamp ciruit isn't biased correctly....the +ve pin of the opamp really ought to be going to 1/2 VCC.

Bill Mountain

Re: Is there something inherently noisy in this design?
« Reply #2 on: May 21, 2012, 01:22:10 PM »
The 10uf wasn’t necessary when I simulated the circuit but I needed it on the breadboard most likely due to running on single supply.

If you are running on a single supply, then that 1st opamp ciruit isn't biased correctly....the +ve pin of the opamp really ought to be going to 1/2 VCC.

True this was just a simple mockup.  You can also leave it in with a dual supply as a HPF.

cpm

Re: Is there something inherently noisy in this design?
« Reply #3 on: May 21, 2012, 06:57:33 PM »
An inverting opamp stage doesnt use that cap to ground. Seems like you are using the topology of a non-inverting stage, which here wont work as expected.
For less noise, you could scale the resistor values down, 10 or even 100 times smaller. Then add buffer at the input to achieve high impedance (a FET buffer is very simple and has low parts count)

Bill Mountain

Re: Is there something inherently noisy in this design?
« Reply #4 on: May 21, 2012, 07:17:37 PM »
An inverting opamp stage doesnt use that cap to ground. Seems like you are using the topology of a non-inverting stage, which here wont work as expected.
For less noise, you could scale the resistor values down, 10 or even 100 times smaller. Then add buffer at the input to achieve high impedance (a FET buffer is very simple and has low parts count)

You need a cap to ground when you have a voltage divider like this on the feedback loop and you're using a single supply.  Opamps can do a lot of crazy things that you don't normally see in pedals.

This circuit was inspired by RG's Adjusticator.  Which has the voltage divider going to vbias instead of an added cap to ground.  Maybe I should try that next time.
« Last Edit: May 21, 2012, 07:25:31 PM by Bill Mountain »

gritz

Re: Is there something inherently noisy in this design?
« Reply #5 on: May 21, 2012, 08:47:41 PM »
My gut feel is you're hearing thermal noise from those large resistances. Remember also that with the gain pot in the "minimum" position the opamp's -ve input is effectively connected to ground via the cap and 3k9 resistor - and it's far too late here to imagine what that does to the circuit's function with those high impedances, beyond it being a zero input of sorts.  :icon_smile:

Gurner

Re: Is there something inherently noisy in this design?
« Reply #6 on: May 22, 2012, 04:25:55 AM »
You need a cap to ground when you have a voltage divider like this on the feedback loop and you're using a single supply.  Opamps can do a lot of crazy things that you don't normally see in pedals.

No, if you are feeding back to the -ve input  - as you are - you don't use a potential divider like that....your potential divider in this case (ie when connecting to the negative pin) is the interaction of the 1M & the 250k feedback resistor (though I think you've reversed the resistors....normally the feedback resistor is larger than the input resistor).

This circuit was inspired by RG's Adjusticator.  Which has the voltage divider going to vbias instead of an added cap to ground.  Maybe I should try that next time.

If you re-check the schem you mentioned, you'll see that there is a potential divider on the +ve pin, but that's not a signal potential divider, but to set the opamp bias correctly (which I mentioned above that in your present guise isn't correct)...and the 10uf used in your reference schem isn't so much a DC blocking cap but more a DC smooting/reservoir cap.
« Last Edit: May 22, 2012, 04:40:08 AM by Gurner »

Bill Mountain

Re: Is there something inherently noisy in this design?
« Reply #7 on: May 22, 2012, 08:30:21 AM »
You need a cap to ground when you have a voltage divider like this on the feedback loop and you're using a single supply.  Opamps can do a lot of crazy things that you don't normally see in pedals.

No, if you are feeding back to the -ve input  - as you are - you don't use a potential divider like that....your potential divider in this case (ie when connecting to the negative pin) is the interaction of the 1M & the 250k feedback resistor (though I think you've reversed the resistors....normally the feedback resistor is larger than the input resistor).

This circuit was inspired by RG's Adjusticator.  Which has the voltage divider going to vbias instead of an added cap to ground.  Maybe I should try that next time.

If you re-check the schem you mentioned, you'll see that there is a potential divider on the +ve pin, but that's not a signal potential divider, but to set the opamp bias correctly (which I mentioned above that in your present guise isn't correct)...and the 10uf used in your reference schem isn't so much a DC blocking cap but more a DC smooting/reservoir cap.


I forgot that the Adjusticator was more than one circuit.  I was referring to the “High Impedance Adjstusticator” here:
http://www.geofex.com/circuits/Adjusticator.gif

I need to remember that not everyone can read my mind.  The drawing was not meant to be a working schematic.  It was just to represent a building block idea.  The grounded non-inverting input should probably be changed to say, “Vbias” to avoid any further confusion but then most of these posts would no longer make sense.

Of course the 10uf is not necessary but it can be used and I did use it so I included it in the drawing.  At the time I couldn’t get it to work without it and that’s because I forgot that in RG’s version the resistor goes to Vbias and not ground.  If I used a smaller cap I could then cut some lows if the stage were too muddy.

I still sometimes use a cap to ground in dual supply operation because of the high pass effect.  It’s the same reason most Tube Screamer and Distortion + variations use a cap to ground instead of eliminating it and connecting to resistor to Vbais instead.  It’s a quick and easier way to avoid farty lows.